U.S. patent application number 13/658557 was filed with the patent office on 2013-05-02 for noise detecting device, noise detecting method, and program.
This patent application is currently assigned to SONY CORPORATION. The applicant listed for this patent is Sony Corporation. Invention is credited to Hiroyasu Baba, Naoki Komine.
Application Number | 20130107059 13/658557 |
Document ID | / |
Family ID | 48172029 |
Filed Date | 2013-05-02 |
United States Patent
Application |
20130107059 |
Kind Code |
A1 |
Baba; Hiroyasu ; et
al. |
May 2, 2013 |
NOISE DETECTING DEVICE, NOISE DETECTING METHOD, AND PROGRAM
Abstract
A noise detecting device includes an image acquiring unit and a
parallel processing unit. The image acquiring unit acquires an
image from a camera that captures the image of a test pattern
including a plurality of lines tilted at a specific angle relative
to a specific direction. The parallel processing unit sequentially
extracts image signals for two lines extending in parallel in the
specific direction of the image, the two parallel lines being away
from each other by a specific number of lines, and performs in a
parallel manner, for each image signals for the two parallel lines,
processing for detecting noise generated in the image on the basis
of a difference in pixel values calculated from the image signals
for the two parallel lines.
Inventors: |
Baba; Hiroyasu; (Kanagawa,
JP) ; Komine; Naoki; (Saitama, JP) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Sony Corporation; |
Tokyo |
|
JP |
|
|
Assignee: |
SONY CORPORATION
Tokyo
JP
|
Family ID: |
48172029 |
Appl. No.: |
13/658557 |
Filed: |
October 23, 2012 |
Current U.S.
Class: |
348/188 ;
348/E17.002 |
Current CPC
Class: |
G06T 5/002 20130101;
G06T 7/0002 20130101; H04N 17/002 20130101 |
Class at
Publication: |
348/188 ;
348/E17.002 |
International
Class: |
H04N 17/00 20060101
H04N017/00 |
Foreign Application Data
Date |
Code |
Application Number |
Oct 28, 2011 |
JP |
2011-237743 |
Claims
1. A noise detecting device comprising: an image acquiring unit
that acquires an image from a camera that captures the image of a
test pattern including a plurality of lines tilted at a specific
angle relative to a specific direction; and a parallel processing
unit that sequentially extracts image signals for two lines
extending in parallel in the specific direction of the image, the
two parallel lines being away from each other by a specific number
of lines, and that performs in a parallel manner, for each image
signals for the two parallel lines, processing for detecting noise
generated in the image on the basis of a difference in pixel values
calculated from the image signals for the two parallel lines.
2. The noise detecting device according to claim 1, wherein the
parallel processing unit includes: a two-line extracting part that
extracts from the image the two parallel lines, which are away from
each other by the specific number of lines, as a reference line and
a test line; a fitting part that calculates the amount of shift in
the specific direction of patterns of pixel values of the reference
line and the test line on the basis of the specific angle of the
tilted test pattern and the number of lines between the reference
line and the test line; a difference calculating part that makes
the patterns of the pixel values match one another on the basis of
the amount of shift to calculate the difference in the pixel values
of the reference line and the test line; and a noise determining
part that determines that noise is not generated when the
difference does not exceed a threshold and that noise is generated
when the difference exceeds the threshold over a specific
range.
3. The noise detecting device according to claim 2, wherein the
test pattern and the camera are relatively moved in the specific
direction, and wherein the two-line extracting part, the fitting
part, the difference calculating part, and the noise determining
part, each of which is arranged in a plural form in parallel,
detect noise in such a manner that a pair of reference line and
test line extracted by one of the two-line extracting parts is not
the same as a pair of reference line and test line extracted by a
different one of the two-line extracting parts.
4. The noise detecting device according to claim 3, wherein the
test pattern and the camera are relatively moved within a certain
range at a constant velocity or at a velocity variable in a
sinusoidal wave manner.
5. The noise detecting device according to claim 4, further
comprising a noise position storing unit that stores a detection
position of noise determined by the noise determining part into a
memory.
6. The noise detecting device according to claim 5, wherein the
test pattern has periodic oblique stripes expressing
characteristics of a pattern of a natural scene, the
characteristics of the pattern of the natural scene including a
high-frequency region, a gradation region, a light same-color
region, and a dark same-color region.
7. A noise detecting method comprising: acquiring an image from a
camera that captures the image of a test pattern including a
plurality of lines tilted at a specific angle relative to a
specific direction; and sequentially extracting image signals for
two lines extending in parallel in the specific direction of the
image, the two parallel lines being away from each other by a
specific number of lines, and performing in a parallel manner, for
each image signals for the two parallel lines, processing for
detecting noise generated in the image on the basis of a difference
in pixel values calculated from the image signals for the two
parallel lines.
8. A program for causing a computer to execute processing
comprising: acquiring an image from a camera that captures the
image of a test pattern including a plurality of lines tilted at a
specific angle relative to a specific direction; and sequentially
extracting image signals for two lines extending in parallel in the
specific direction of the image, the two parallel lines being away
from each other by a specific number of lines, and performing in a
parallel manner, for each image signals for the two parallel lines,
processing for detecting noise generated in the image on the basis
of a difference in pixel values calculated from the image signals
for the two parallel lines.
Description
CROSS REFERENCES TO RELATED APPLICATIONS
[0001] The present application claims priority to Japanese Priority
Patent Application JP 2011-237743 filed in the Japan Patent Office
on Oct. 28, 2011, the entire content of which is hereby
incorporated by reference.
BACKGROUND
[0002] The present disclosure relates to a noise detecting device,
a noise detecting method, and a program that detect, for example,
noise superimposed on moving images captured with a camera as a
test object.
[0003] Due to the development of manufacturing technology of
cameras, it has become possible to easily capture moving images.
Cameras often capture images of natural materials. Images of
natural materials contain various components such as, for example,
frequency components. In the explanation given below, scenes
photographed with cameras will be referred to as "natural
scenes".
[0004] FIG. 13 illustrates an example of an image of a natural
scene captured with a camera.
[0005] The image illustrated in FIG. 13 contains buildings and the
sky. The image includes a "gradation region" representing changes
in the color tone of the sky and a "same-color region" representing
a certain area having the same color. The image also includes a
"high-frequency region" representing a repetitive pattern, such as
a fine pattern or the like, and an "edge region" corresponding to
corners of buildings and the like and representing a portion
exhibiting a high contrast ratio.
[0006] In order to capture the image of the natural scene
illustrated in FIG. 13 as a natural moving image, various
performance tests are performed on cameras.
[0007] FIG. 14 illustrates an example of the configuration of an
imaging test system 100 of related art that performs a test for
noise superimposed on moving images.
[0008] The imaging test system 100 includes a natural scene 101, a
camera 102 that captures an image of the natural scene 101 as a
moving image and for which a test as to whether noise is
superimposed on the moving image is performed, and a test monitor
103 that displays the moving image output from the camera 102. The
case where the camera 102 is used as a test object so that the
imaging performance of the camera 102 is tested is illustrated as
the example of FIG. 14. The camera 102 is irradiated with strong
radio waves on the assumption of various disturbances, and the
output level of the power source voltage to be supplied to the
camera 102 is changed. A tester 104 visually observes the moving
image displayed on the test monitor 103 to determine whether or not
noise (for example, block noise, a beat by which an image is
deviated in a horizontal direction, etc.) is superimposed on the
moving image.
[0009] For example, a technology disclosed in Japanese Unexamined
Patent Application Publication No. 2006-325122 is available as a
technology to be used for estimation of noise superimposed on
moving images. In the technology disclosed in Japanese Unexamined
Patent Application Publication No. 2006-325122, a test image is
scrolled through, a sample image that is similar to the test image
is specified, and the index value of the sample image is defined as
the index value of the test image.
SUMMARY
[0010] In the case of visual observation tests carried out by a
plurality of testers 104, the skills of determining whether or not
noise is generated vary from individual to individual. Furthermore,
since the tester 104 stays in front of the test monitor 103 for a
long time for a test that relies on human operation, the
concentration of the tester 104 may not last long, thus causing
false test results.
[0011] In addition, in the case of a moving image obtained by
capturing an image of a natural scene, even when a test is
automated, it is difficult to determine whether a portion
determined by the tester 104 to be noise is actual noise or a
pattern existing in the natural scene. Furthermore, even with the
technology disclosed in Japanese Unexamined Patent Application
Publication No. 2006-325122, the level of the skill of comparing a
sample image with a test image differs depending on the ability of
the tester 104. Thus, a plurality of testers 104 may produce
different test results. In addition, in the case where a moving
image used as a reference for a noise test is stored in a memory,
since a large memory capacity is used for such a moving image, a
limited amount of memory capacity allocated to a noise detecting
device is not efficiently used. Furthermore, in order to achieve
synchronization between a reference moving image and a moving image
to be tested, a new mechanism for achieving synchronization timing
is used, thus taking more time, effort, and money.
[0012] It is desirable to efficiently perform a test for noise
superimposed on a moving image.
[0013] According to an embodiment of the present disclosure, an
image is acquired from a camera that captures the image of a test
pattern including a plurality of lines tilted at a specific angle
relative to a specific direction.
[0014] Image signals for two lines extending in parallel in the
specific direction of the image, the two parallel lines being away
from each other by a specific number of lines, are sequentially
extracted, and processing for detecting noise generated in the
image is performed in a parallel manner, for each image signals for
the two parallel lines, on the basis of the difference in pixel
values calculated from the image signals for the two parallel
lines.
[0015] Accordingly, with the use of a difference in image signals
for two lines extending in parallel in the specific direction,
processing for detecting noise generated in an image can be
performed automatically.
[0016] According to an embodiment of the present disclosure, an
image acquired from a camera includes a test pattern including a
plurality of lines tilted at a specific angle relative to a
specific direction. Image signals for two lines extending in
parallel in the specific direction of the image are extracted, and
a portion where a difference in pixel values is generated is
determined to be a position where noise is generated. Thus, noise
determination can be automatically achieved, irrespective of the
skill of a tester.
[0017] Additional features and advantages are described herein, and
will be apparent from the following Detailed Description and the
figures.
BRIEF DESCRIPTION OF THE FIGURES
[0018] FIG. 1 is a perspective view illustrating an example of the
external configuration of a moving image noise detecting system
according to an embodiment of the present disclosure;
[0019] FIG. 2 is a block diagram illustrating an example of the
internal configuration of a moving image noise detecting device
according to an embodiment of the present disclosure;
[0020] FIG. 3 is an explanatory diagram illustrating an example of
the conceptual configuration of a test pattern used in an
embodiment of the present disclosure;
[0021] FIG. 4 is an explanatory diagram illustrating an example of
a specific configuration of a test pattern used in an embodiment of
the present disclosure;
[0022] FIG. 5 is an explanatory diagram illustrating an example of
the operation of a test pattern used in an embodiment of the
present disclosure;
[0023] FIG. 6 is an explanatory diagram illustrating an example of
a moving image of a test pattern for one frame in an embodiment of
the present disclosure;
[0024] FIG. 7 is an explanatory diagram illustrating an example of
a reference line and a test line in the case where a specific test
pattern is used in an embodiment of the present disclosure;
[0025] FIGS. 8A and 8B are explanatory diagrams of the patterns of
pixel values of a reference line and a test line, respectively, in
an embodiment of the present disclosure;
[0026] FIGS. 9A and 9B are explanatory diagrams illustrating
examples of the patterns of pixel values of a reference line and a
test line, respectively, in an embodiment of the present
disclosure;
[0027] FIGS. 10A and 10B are explanatory diagrams illustrating
examples of the patterns of pixel values in the case where noise is
superimposed on a reference line in an embodiment of the present
disclosure;
[0028] FIG. 11 is an explanatory diagram illustrating an example in
which differences in the pixel values of a reference line and a
test line that are made to match one another in the vertical
direction are plotted in an embodiment of the present
disclosure;
[0029] FIGS. 12A and 12B are explanatory diagrams illustrating an
image in which noise is superimposed and an example in which a
detection result of the noise is illustrated, respectively, in an
embodiment of the present disclosure;
[0030] FIG. 13 is an explanatory diagram illustrating an example of
an image of a natural scene captured with a camera; and
[0031] FIG. 14 illustrates the external configuration of an imaging
test system of related art that performs a test for noise
superimposed on a moving image.
DETAILED DESCRIPTION
[0032] Hereinafter, embodiments of the present disclosure will be
described. The description will be given in the following
order:
[0033] 1. Embodiment (noise detection: example of processing for
determining whether noise is generated on the basis of two
horizontal lines)
[0034] 2. Modifications
1. Embodiment (Example of Processing for Determining Whether Noise
is Generated on the Basis of Two Horizontal Lines)
[0035] Hereinafter, an embodiment of the present disclosure
(hereinafter, referred to as "this embodiment") will be explained
with reference to FIGS. 1 to 12B. In this embodiment, an example
will be explained in which this embodiment is applied to a moving
image noise detecting device 1 that employs a moving image noise
detecting method for automatically detecting noise superimposed on
a moving image captured with a camera. First, an example of the
configuration of the moving image noise detecting device 1 will be
explained.
[0036] FIG. 1 illustrates an example of the external configuration
of a moving image noise detecting system 10 according to this
embodiment.
[0037] The moving image noise detecting system 10 includes the
moving image noise detecting device 1 that detects noise
superimposed on a moving image. The moving image noise detecting
system 10 also includes a test object 2 that is connected to the
moving image noise detecting device 1 and for which a test is
performed for noise superimposed on a captured and output moving
image. The moving image noise detecting system 10 also includes a
plate-like test panel 4 on which a test pattern is printed. In this
embodiment, a video camera that is capable of capturing moving
images is used as the test object 2.
[0038] The test pattern printed on the test panel 4 is a striped
pattern including a plurality of lines that are tilted at a
specific angle relative to a specific direction, as described
below. The test panel 4 is relatively moved in a horizontal
direction at a uniform velocity by a pattern operation unit 3. In
this embodiment, "a horizontal direction" is defined as the
specific direction, and "an angle of 45 degrees in the upper right
direction relative to the horizontal direction" is defined as the
specific angle of the test pattern. The test object 2 captures an
image of the moving test pattern, and the moving image noise
detecting device 1 performs processing for detecting noise from the
image captured with the test object 2, so that the performance of
the test object 2 can be estimated. The moving image noise
detecting device 1 and the pattern operation unit 3 may be
connected to each other so that the moving velocity and position of
the pattern operation unit 3 can be changed in accordance with an
instruction from the moving image noise detecting device 1.
[0039] FIG. 2 illustrates an example of the internal configuration
of the moving image noise detecting device 1.
[0040] The moving image noise detecting device 1 includes an image
acquiring unit 11 that acquires a moving image obtained by
capturing an image of the test pattern from the test object 2
serving as a test object. The moving image noise detecting device 1
also includes a parallel processing unit 12 that performs specific
processing in a parallel manner for each two lines extracted from
the moving image. The moving image noise detecting device 1 also
includes a noise position storing unit 13 that stores the position
of noise in the moving image determined by the parallel processing
unit 12 into a memory 14. The memory 14 may be, for example, a
rewritable random access memory (RAM).
[0041] The parallel processing unit 12 performs processing for
extracting image signals for two horizontal lines extending in
parallel in the horizontal scanning direction of a moving image,
the two parallel horizontal lines being away from each other by a
specific number of lines, and detecting, for each image signals for
the two parallel horizontal lines in a parallel manner, noise
generated in the moving image on the basis of a difference in pixel
values obtained from the image signals for the two parallel
horizontal lines.
[0042] The parallel processing unit 12 includes a two-line
extracting part 15-1 that extracts image signals for two horizontal
lines, which are away from each other by a specific number of lines
(.DELTA.n lines), from a moving image for one frame acquired by the
image acquiring unit 11. One of the two horizontal lines is defined
as a reference line and the other horizontal line is defined as a
test line.
[0043] The parallel processing unit 12 also includes a fitting part
16-1 that calculates the amount of shift in the horizontal
direction of the pattern of individual pixel values on the basis of
a specific angle of the tilt of the test pattern relative to the
horizontal direction and the number of lines between the reference
line and the test line. The fitting part 16-1 calculates, as the
amount of shift in the horizontal direction, the minimum sum of
squares of the difference between the pixel values of the reference
line and the test line.
[0044] The parallel processing unit 12 also includes a difference
calculating part 17-1 that calculates the difference in the pixel
values of the reference line and the test line by making the
patterns of the individual pixel values match one another on the
basis of the calculated amount of shift in the horizontal
direction. The parallel processing unit 12 also includes a noise
determining part 18-1 that determines that noise is not generated
when the difference does not exceed a threshold and determines that
noise is generated when the difference exceeds the threshold over a
certain range.
[0045] As described above, the two-line extracting part 15-1 to the
noise determining part 18-1 determine whether noise is generated in
image signals for two lines that are away from each other by
.DELTA.n lines. In order to determine whether noise is generated in
image signals for the other lines, a number (n) of parallel
processing operations corresponding to half the number of vertical
lines of the imaging area of the test object 2 are performed. Thus,
although all the processing parts are not illustrated, a plurality
of sets of two-line extracting parts 15-1 to 15-n, fitting parts
16-1 to 16-n, difference calculating parts 17-1 to 17-n, and noise
determining parts 18-1 to 18-n that are arranged in parallel detect
noise in a parallel manner. A pair of reference line and test line
that are extracted by a two-line extracting part is not the same as
a pair of reference line and test line that are extracted by a
different two-line extracting part. For example, in the case where
the number of horizontal pixels of the imaging area of the test
object 2 is 1920 and the number of vertical lines of the imaging
area of the test object 2 is 1080, n represents 540 (=1080/2).
[0046] An example of the configuration of a test pattern will now
be explained with reference to FIGS. 3 and 4.
[0047] FIG. 3 illustrates an example of the conceptual
configuration of a test pattern.
[0048] The test pattern is a periodic pattern that expresses
characteristics existing in a natural scene, such as a
high-frequency region a, a gradation region b, a same-color region
(dark) c, a same-color region (light) d, and the like, using a
plurality of consecutive lines.
[0049] As described above, the test pattern has stripes including a
plurality of oblique lines that are tilted at, as a specific angle,
an angle of 45 degrees, in the upper right direction. With the use
of this test pattern, the same processing can be applied to any
horizontal scanning period in a moving image captured with the test
object 2. This test pattern is suitable for increasing the speed of
a noise test. Furthermore, with the oblique lines, when the test
pattern is moved only in the horizontal direction, images can be
captured as if the stripes were moved in horizontal and vertical
directions.
[0050] FIG. 4 illustrates an example of a specific configuration of
the test pattern.
[0051] The test pattern used in this embodiment has stripes of
black and white. By changing the color depth of the stripes,
unifying the color, or making the widths of the stripes narrower,
the stripes simulate various components appearing in a natural
scene.
[0052] For example, in the gradation region, the brightness
continuously changes between black and white. The high-frequency
region has a plurality of oblique lines that are narrower than the
others. The dark same-color region has one wide black oblique line,
and the light same-color region has one wide white oblique line. A
portion where the dark same-color region and the light same-color
region are adjacent to each other is defined as an edge region
exhibiting a high contrast ratio.
[0053] An example of the operation of the test pattern will now be
explained with reference to FIGS. 5 to 11.
[0054] FIG. 5 illustrates an example of the operation of the test
pattern.
[0055] The test pattern is moved in the horizontal direction by the
pattern operation unit 3. The pattern operation unit 3 moves the
test panel 4 at a constant velocity in such a manner that an
imaging area 5 of the test object 2 is located inside the test
pattern. The pattern operation unit 3 may move the test panel 4
with a variable movement in a sinusoidal wave manner, in which the
test panel 4 is moved slowly at first, then the movement speed
gradually increases, and the test panel 4 is gradually stopped at
the end. A moving image for one frame of the test pattern captured
with the test object 2 is acquired by the image acquiring unit
11.
[0056] FIG. 6 illustrates an example of a moving image of the test
pattern for one frame.
[0057] The two-line extracting part 15-1 performs horizontal
scanning of a moving image acquired by the image acquiring unit 11
and obtains a reference line 6 and a test line 7. A horizontal line
that is away from the reference line 6, which is represented by a
solid line in the moving image for one frame, by .DELTA.n lines is
defined as the test line 7. The two lines extracted at this time,
that is, the reference line 6 and the test line 7, are extracted in
such a manner that all the lines in the image for one frame are
classified into either the reference line 6 or the test line 7. For
example, the two-line extracting parts 15-1 to 15-n extract all the
odd-numbered horizontal lines as the reference lines 6 and extract
all the even-numbered horizontal lines as the test lines 7.
Accordingly, all the odd-numbered lines included in an image for
one frame are defined as the reference lines 6.
[0058] FIG. 7 illustrates an example of the reference line 6 and
the test line 7 in the case where a specific test pattern is
used.
[0059] The two-line extracting parts 15-1 to 15-n extract the
reference lines 6 and the test lines 7 from the test pattern
including the various regions described above. Thus, the individual
lines have dark portions and light portions. It is assumed that a
graph representing the pixel values of each horizontal line
exhibits a constant pattern.
[0060] FIGS. 8A and 8B illustrate examples of the patterns of pixel
values of the reference line 6 and the test line 7. FIG. 8A
illustrates an example of the pattern of the pixel values of the
reference line 6 (the nth line). FIG. 8B illustrates an example of
the pattern of the pixel values of the test line 7 (the
n+.DELTA.nth line).
[0061] In the patterns of the pixel values illustrated in FIGS. 8A
and 8B, two lines are represented by pixel values using 256
grayscale levels, where "0" represents black and "255" represents
white. As described above, the two lines each include the
same-color region (light), the same-color region (dark), the
high-frequency region, the gradation region, and the edge region.
Since the test pattern is tilted at a specific angle, the pattern
of the pixel values of the two lines may slightly differ from each
other in the horizontal direction.
[0062] An example in which noise superimposed on the pattern of
pixel values is detected will now be explained.
[0063] FIGS. 9A to 10B illustrate examples of the patterns of pixel
values in the case where noise is superimposed on the reference
line 6. FIG. 9A illustrates an example of the pattern of the pixel
values of the reference line. FIG. 9B illustrates an example of the
pattern of the pixel values of the test line. FIG. 10A illustrates
an example of the pattern of the pixel values of the reference
line. FIG. 10B illustrates an example of the pattern of the pixel
values of the test line. As illustrated in FIG. 10A, block noise is
generated in the gradation region of the reference line 6.
[0064] As described above, the patterns of the pixel values of the
two lines differ from each other in the horizontal direction by
.DELTA.x pixels even for a moving image captured for the same
frame. Thus, the fitting parts 16-1 to 16-n calculate the amount
.DELTA.x of shift, which is the minimum sum of squares of the
difference between the patterns of the pixel values of
corresponding pixels of the two lines. Since the theoretical value
of the value .DELTA.x can be calculated using equation (1) provided
below on the basis of the tilt .theta. of the pattern and the value
.DELTA.n, the sums of squares for positions around the theoretical
value are calculated, so that an accurate value .DELTA.x can be
obtained quickly.
.DELTA.x=.DELTA.n tan .theta. (1)
[0065] Then, the fitting parts 16-1 to 16-n perform fitting
processing for shifting the test line 7 by .DELTA.x to make the
positions in the horizontal direction represented by the patterns
of the pixel values of the reference line 6 and the test line 7
match one another.
[0066] FIG. 11 illustrates an example of plotted differences in the
pixels values of a reference line and a test line that are made to
match one another in the vertical direction.
[0067] The difference calculating parts 17-1 to 17-n perform
calculation for obtaining differences in the pixel values of the
pixel patterns of two lines at the positions that are made to match
one another by the fitting parts 16-1 to 16-n. In the case of two
lines for which noise is not detected in accordance with the result
of the calculation of a difference, an ideal difference is 0.
Meanwhile, when noise is detected, the difference is not 0.
[0068] The noise determining parts 18-1 to 18-n calculate local
averages of differences. When an average exceeds a set threshold,
it is determined that noise is generated at the corresponding
position. The local average of differences is calculated using
equation (2) provided below, where "j" represents the initial
position at which the local average of differences is calculated,
and ".DELTA.j" represents the length over which the average is
taken.
D Ave = 1 .DELTA. j k = j j + .DELTA. j D ( k ) ( 2 )
##EQU00001##
[0069] As described above, the noise determining parts 18-1 to 18-n
detect generation of noise at a specific position of a specific
line. The noise position storing unit 13 collects information on
the positions of noise determined by the noise determining parts
18-1 to 18-n for individual two lines of all the horizontal lines,
and stores the positions of noise in the image for one frame into
the memory 14.
[0070] After acquiring an image and storing the position where
noise is generated, the moving image noise detecting device 1
acquires an image for the next frame from an object and performs
similar processing. Then, the moving image noise detecting device 1
performs the above-described series of processing operations for
all the frames in real time to achieve noise determination at high
speed.
[0071] The moving image noise detecting device 1 according to this
embodiment described above determines whether noise is generated by
extracting the reference line 6 and the test line 7 from all the
lines within one frame and performing parallel processing for the
number of pairs, which corresponds to approximately half the
direction of the height of an image. As described above, since the
same processing can be applied to all the horizontal lines included
in a moving image, the noise detecting method according to this
embodiment is suitable for parallel processing. Therefore, a test
can be performed thoroughly over an image, and noise detection can
be performed over a moving image at high speed without any
horizontal lines for which a test is not performed.
[0072] In addition, since the comparison between the pixel values
of two horizontal lines included in an image is performed in real
time, a memory for only one line is provided to store the reference
line 6 and a memory for storing the test line 7 is not provided.
Thus, the amount of memory capacity to be used can be reduced, and
the cost of the memory to be implemented in the moving image noise
detecting device 1 can be reduced.
[0073] In addition, since the positions of the reference line 6 and
the test line 7 that are away from each other by .DELTA.n lines are
made to match one another in the vertical direction and the
comparison between the pixel values of the reference line 6 and the
test line 7 is performed, beat noise as well as block noise can be
easily detected. Since the positions of individual lines in the
vertical direction can be made to match one another in units of
pixels, a very small beat noise, which is difficult to be viewed by
the eyes of a tester, can be detected.
[0074] In addition, the test pattern has stripes including oblique
lines. Thus, even when an image of the test pattern moving in the
horizontal direction is captured with a camera at a fixed position,
images can be obtained as if the stripes were moved vertically
above or below depending on the moving direction of the test
pattern.
2. Modifications
[0075] When noise detecting processing is not performed in real
time, the processing may be speeded up by eliminating a number of
horizontal lines to be tested. Even in this case, block noise or
the like superimposed on an image for a plurality of lines can be
detected.
[0076] In the embodiment described above, a noise test is performed
using a test pattern on which stripes including consecutive black
and white lines are printed. However, the stripes may be in color.
With the use of a test pattern on which stripes in color are
printed, reproduction of color tone and the like can also be
tested.
[0077] In addition, a similar noise test can be performed for still
images as well as moving images. Thus, not only a video camera but
also a still camera may be used as the test object 2. Furthermore,
the number of horizontal lines to be extracted in a single
processing operation is not limited to two. Noise detection may be
performed using three or more extracted horizontal lines.
[0078] In addition, in the fitting processing performed by the
fitting parts 16-1 to 16-n, the amount .DELTA.x of shift, which is
the minimum sum of squares of the difference between the patterns
of pixel values, is calculated. However, the amount .DELTA.x may be
calculated in a different way. For example, the amount .DELTA.x may
be calculated using the average of differences between the patterns
of pixel values.
[0079] In addition, the moving image noise detecting device 1 may
extract image signals for two parallel vertical lines, instead of
two parallel horizontal lines. Furthermore, two lines in any
direction may be extracted as long as the two lines extend in
parallel.
[0080] In addition, the angle of the tilt of the test pattern
relative to a horizontal line may not be set to an angle of 45
degrees. Any angle and moving direction may be set as long as the
test pattern is tilted at a specific angle in a specific direction
in advance and the angle and the relative moving direction of the
test pattern and the camera are set in advance.
[0081] The series of processing operations according to the
above-described embodiments can be executed by hardware or
software. When the series of processing operations are executed by
software, the series of processing operations can be executed by a
computer in which a program forming the software is built in
dedicated hardware or a computer to which a program for
implementing various functions is installed. For example, a program
forming desired software may be installed to a general-purpose
personal computer or the like.
[0082] A recording medium recording a program code of the software
for implementing the functions of the above-described embodiments
may be supplied to a system or an apparatus. Obviously, the
functions can also be implemented when a computer (or a controller
such as a central processing unit (CPU)) of the system or the
apparatus reads and executes the program code stored on the
recording medium.
[0083] As the recording medium for supplying the program code in
this case, for example, a flexible disk, a hard disk, an optical
disk, a magneto-optical disk, a compact disc read-only memory
(CD-ROM), a compact disc readable (CD-R), a magnetic tape, a
nonvolatile memory card, or a ROM may be used.
[0084] The functions of the above-described embodiments may be
implemented when the computer executes the read program code.
Furthermore, an operating system (OS) or the like running on the
computer executes part or all of the actual processing on the basis
of instructions of the program code. The case where the functions
of the above-described embodiments are implemented by the
processing may also be included in the present disclosure.
[0085] It is obvious that the present disclosure is not limited to
the above-described embodiments and various other applications and
modifications may be made to the present disclosure without
departing from the spirit of the present disclosure.
[0086] The present disclosure may employ the configurations
described below.
[0087] (1) A noise detecting device including
[0088] an image acquiring unit that acquires an image from a camera
that captures the image of a test pattern including a plurality of
lines tilted at a specific angle relative to a specific direction,
and
[0089] a parallel processing unit that sequentially extracts image
signals for two lines extending in parallel in the specific
direction of the image, the two parallel lines being away from each
other by a specific number of lines, and that performs in a
parallel manner, for each image signals for the two parallel lines,
processing for detecting noise generated in the image on the basis
of a difference in pixel values calculated from the image signals
for the two parallel lines.
[0090] (2) The noise detecting device described in (1), wherein the
parallel processing unit includes
[0091] a two-line extracting part that extracts from the image the
two parallel lines, which are away from each other by the specific
number of lines, as a reference line and a test line,
[0092] a fitting part that calculates the amount of shift in the
specific direction of patterns of pixel values of the reference
line and the test line on the basis of the specific angle of the
tilted test pattern and the number of lines between the reference
line and the test line,
[0093] a difference calculating part that makes the patterns of the
pixel values match one another on the basis of the amount of shift
to calculate the difference in the pixel values of the reference
line and the test line, and
[0094] a noise determining part that determines that noise is not
generated when the difference does not exceed a threshold and that
noise is generated when the difference exceeds the threshold over a
specific range.
[0095] (3) The noise detecting device described in (1) or (2),
[0096] wherein the test pattern and the camera are relatively moved
in the specific direction, and
[0097] wherein the two line extracting part, the fitting part, the
difference calculating part, and the noise determining part, each
of which is arranged in a plural form in parallel, detect noise in
such a manner that a pair of reference line and test line extracted
by one of the two-line extracting parts is not the same as a pair
of reference line and test line extracted by a different one of the
two-line extracting parts.
[0098] (4) The noise detecting device described in any one of (1)
to (3), wherein the test pattern and the camera are relatively
moved within a certain range at a constant velocity or at a
velocity variable in a sinusoidal wave manner.
[0099] (5) The noise detecting device described in any one of (1)
to (4) further including a noise position storing unit that stores
a detection position of noise determined by the noise determining
part into a memory.
[0100] (6) The noise detecting device described in any one of (1)
to (5), wherein the test pattern has periodic oblique stripes
expressing characteristics of a pattern of a natural scene, the
characteristics of the pattern of the natural scene including a
high-frequency region, a gradation region, a light same-color
region, and a dark same-color region.
[0101] (7) A noise detecting method including
[0102] acquiring an image from a camera that captures the image of
a test pattern including a plurality of lines tilted at a specific
angle relative to a specific direction, and
[0103] sequentially extracting image signals for two lines
extending in parallel in the specific direction of the image, the
two parallel lines being away from each other by a specific number
of lines, and performing in a parallel manner, for each image
signals for the two parallel lines, processing for detecting noise
generated in the image on the basis of a difference in pixel values
calculated from the image signals for the two parallel lines.
[0104] (8) A program for causing a computer to execute processing
including
[0105] acquiring an image from a camera that captures the image of
a test pattern including a plurality of lines tilted at a specific
angle relative to a specific direction, and
[0106] sequentially extracting image signals for two lines
extending in parallel in the specific direction of the image, the
two parallel lines being away from each other by a specific number
of lines, and performing in a parallel manner, for each image
signals for the two parallel lines, processing for detecting noise
generated in the image on the basis of a difference in pixel values
calculated from the image signals for the two parallel lines.
[0107] It should be understood that various changes and
modifications to the presently preferred embodiments described
herein will be apparent to those skilled in the art. Such changes
and modifications can be made without departing from the spirit and
scope of the present subject matter and without diminishing its
intended advantages. It is therefore intended that such changes and
modifications be covered by the appended claims.
* * * * *